Publications

@article{Yuca2026,
  title = {A {{Survey}} on {{Privacy-Preserving Computing}} in the {{Automotive Domain}}},
  author = {Yuca, Nergiz and Matyunin, Nikolay and Arzoglou, Ektor and Anagnostopoulos, Nikolaos Athanasios and Katzenbeisser, Stefan},
  year = {2026},
  month = feb,
  journal = {ACM Computing Surveys},
  volume = {58},
  number = {5},
  pages = {128:1--128:36},
  issn = {0360-0300},
  doi = {10.1145/3770580},
  abstract = {As vehicles become increasingly connected and autonomous, they accumulate and manage various personal data, thereby presenting a key challenge in preserving privacy during data sharing and processing. This survey reviews applications of Secure Multi-Party Computation (MPC) and Homomorphic Encryption (HE) that address these privacy concerns in the automotive domain. First, we identify the scope of privacy-sensitive use cases for these technologies, by surveying existing works that address privacy issues in different automotive contexts, such as location-based services, mobility infrastructures, traffic management, etc. Then, we review recent works that employ MPC and HE as solutions for these use cases in detail. Our survey highlights the applicability of these privacy-preserving technologies in the automotive domain, while also identifying challenges and gaps in the current research landscape. This work aims to provide a clear and comprehensive overview of this emerging field and to encourage further research in this domain.},
  pdf = {https://arxiv.org/pdf/2508.01798}
}

@inproceedings{Yuca2026a,
  title = {{{Privacy-Preserving}} {{Federated Loss Exploration}} for {{Vision-Based Perception}} under {{Non-IID Data}}},
  booktitle = {IEEE International Conference on Intelligent Transportation Systems (ITSC), to appear},
  author = {Yuca, Nergiz and Intern{\`o}, Christian and Matyunin, Nikolay and Olhofer, Markus and Hammer, Barbara and Katzenbeisser, Stefan},
  year = {2026},
  publisher = {IEEE},
  abstract = {Vision-based image classification models benefit from Federated Learning (FL) across vehicles, however learning accurate models from distributed vehicular cameras is challenging due to data heterogeneity and privacy concerns. To improve performance on non-IID data, advanced FL optimization approaches have emerged that share auxiliary insights, such as client gradient behaviors. While these gradient-guided optimized FL methods improve the model accuracy, their reliance on sharing additional gradient information introduces a previously unexplored privacy vulnerability. In our empirical privacy-leakage experiment, we show that a malicious server can infer clients' label sets from shared guidance matrices under heterogeneous client label distributions (simulated via Dirichlet-based label partitions), between 50% and 84% accuracy. To address this privacy leakage, we propose PPFLex that replaces the central server with three Secure Multi-Party Computation (MPC) servers to securely compute the global guidance matrix and perform federated averaging under semi-honest and malicious MPC server assumptions. We evaluate PPFLex on widely used traffic-sign classification datasets (e.g., GTSRB and BTSC) as representative autonomous vehicle perception benchmarks, comparing against baselines. Our results show that PPFLex achieves accuracy comparable to the plain-text FedLEx pipeline while preventing privacy leakage, incurring acceptable MPC overhead.},
  pdf = {https://openreview.net/pdf?id=PTdCBanvv3}
}

@article{Raza2026,
  title = {Amnesia: {{Adversarial Semantic Layer Specific Activation Steering}} in {{Large Language Models}}},
  shorttitle = {Amnesia},
  author = {Raza, Ali and Gupta, Gaurang and Matyunin, Nikolay and Patra, Jibesh},
  year = {2026},
  month = mar,
  journal = {arXiv preprint arXiv:2603.10080},
  eprint = {2603.10080},
  archiveprefix = {arXiv},
  abstract = {Large Language Models (LLMs) have the potential to create harmful content, such as generating sophisticated phishing emails and assisting in writing code of harmful computer viruses. Thus, it is crucial to ensure their safe and responsible response generation. To reduce the risk of generating harmful or irresponsible content, researchers have developed techniques such as reinforcement learning with human feedback to align LLM's outputs with human values and preferences. However, it is still undetermined whether such measures are sufficient to prevent LLMs from generating interesting responses. In this study, we propose Amnesia, a lightweight activation-space adversarial attack that manipulates internal transformer states to bypass existing safety mechanisms in open-weight LLMs. Through experimental analysis on state-of-the-art, open-weight LLMs, we demonstrate that our attack effectively circumvents existing safeguards, enabling the generation of harmful content without the need for any fine-tuning or additional training. Our experiments on benchmark datasets show that the proposed attack can induce various antisocial behaviors in LLMs. These findings highlight the urgent need for more robust security measures in open-weight LLMs and underscore the importance of continued research to prevent their potential misuse.},
  pdf = {https://arxiv.org/pdf/2603.10080}
}

@article{Yuan2026,
  title = {{{LADFA}}: {{A Framework}} of {{Using Large Language Models}} and {{Retrieval-Augmented Generation}} for {{Personal Data Flow Analysis}} in {{Privacy Policies}}},
  shorttitle = {{{LADFA}}},
  author = {Yuan, Haiyue and Matyunin, Nikolay and Raza, Ali and Li, Shujun},
  year = {2026},
  month = jan,
  journal = {arXiv preprint arXiv:2601.10413},
  eprint = {2601.10413},
  archiveprefix = {arXiv},
  abstract = {Privacy policies help inform people about organisations' personal data processing practices, covering different aspects such as data collection, data storage, and sharing of personal data with third parties. Privacy policies are often difficult for people to fully comprehend due to the lengthy and complex legal language used and inconsistent practices across different sectors and organisations. To help conduct automated and large-scale analyses of privacy policies, many researchers have studied applications of machine learning and natural language processing techniques, including large language models (LLMs). While a limited number of prior studies utilised LLMs for extracting personal data flows from privacy policies, our approach builds on this line of work by combining LLMs with retrieval-augmented generation (RAG) and a customised knowledge base derived from existing studies. This paper presents the development of LADFA, an end-to-end computational framework, which can process unstructured text in a given privacy policy, extract personal data flows and construct a personal data flow graph, and conduct analysis of the data flow graph to facilitate insight discovery. The framework consists of a pre-processor, an LLM-based processor, and a data flow post-processor. We demonstrated and validated the effectiveness and accuracy of the proposed approach by conducting a case study that involved examining ten selected privacy policies from the automotive industry. Moreover, it is worth noting that LADFA is designed to be flexible and customisable, making it suitable for a range of text-based analysis tasks beyond privacy policy analysis.},
  pdf = {https://arxiv.org/pdf/2601.10413}
}

@inproceedings{Lauser2025,
  title = {Oblivious {{Plug}}\&{{Charge}}: {{A Privacy-Preserving EV Charging Scheme}} Based on {{ORAM}}},
  shorttitle = {Oblivious {{Plug}}\&{{Charge}}},
  booktitle = {3rd {{USENIX Symposium}} on {{Vehicle Security}} and {{Privacy}} ({{VehicleSec}} 25)},
  author = {Lauser, Timm and Yuca, Nergiz and Kern, Dustin and Matyunin, Nikolay and Katzenbeisser, Stefan and Krau{\ss}, Christoph},
  year = {2025},
  month = aug,
  pages = {221--232},
  publisher = {USENIX Association},
  address = {Seattle, WA, USA},
  isbn = {978-1-939133-49-6},
  abstract = {In the rapidly developing Electric Vehicle (EV) charging infrastructure, ensuring user privacy during charging and billing processes has become a concern. Modern vehicles increasingly support Plug-and-Charge (PnC), a standard for direct authentication and billing of charging sessions without user interaction. However, within the PnC architecture, the exchange of sensitive data enables the e-Mobility service providers and charging station operators to build profiles of users' travel patterns and habits. We introduce a novel Oblivious Random Access Machine (ORAM)-based privacy-preserving database to aggregate the consumption of the EVs at the charging station operator. In our scheme, the charging station operator cannot link individual charging sessions to specific vehicles, while the service provider only obtains the aggregated consumption over the entire billing period instead of details about individual charging sessions.},
  pdf = {https://www.usenix.org/system/files/vehiclesec25-lauser.pdf}
}

@article{Shi2024,
  title = {Uncovering {{CWE-CVE-CPE Relations}} with {{Threat Knowledge Graphs}}},
  author = {Shi, Zhenpeng and Matyunin, Nikolay and Graffi, Kalman and Starobinski, David},
  year = {2024},
  month = feb,
  journal = {ACM Transactions on Privacy and Security},
  volume = {27},
  number = {1},
  pages = {13:1--13:26},
  issn = {2471-2566},
  doi = {10.1145/3641819},
  abstract = {Security assessment relies on public information about products, vulnerabilities, and weaknesses. So far, databases in these categories have rarely been analyzed in combination. Yet, doing so could help predict unreported vulnerabilities and identify common threat patterns. In this article, we propose a methodology for producing and optimizing a knowledge graph that aggregates knowledge from common threat databases (CVE, CWE, and CPE). We apply the threat knowledge graph to predict associations between threat databases, specifically between products, vulnerabilities, and weaknesses. We evaluate the prediction performance both in closed world with associations from the knowledge graph and in open world with associations revealed afterward. Using rank-based metrics (i.e., Mean Rank, Mean Reciprocal Rank, and Hits@N scores), we demonstrate the ability of the threat knowledge graph to uncover many associations that are currently unknown but will be revealed in the future, which remains useful over different time periods. We propose approaches to optimize the knowledge graph and show that they indeed help in further uncovering associations. We have made the artifacts of our work publicly available.},
  pdf = {Shi2024 - Uncovering CWE-CVE-CPE Relations.pdf}
}

@inproceedings{Yuan2024,
  title = {A {{Graph-Based Model}} for {{Vehicle-Centric Data Sharing Ecosystem}}},
  booktitle = {2024 {{IEEE}} 27th {{International Conference}} on {{Intelligent Transportation Systems}} ({{ITSC}})},
  author = {Yuan, Haiyue and Raza, Ali and Matyunin, Nikolay and Patra, Jibesh and Li, Shujun},
  year = {2024},
  month = sep,
  pages = {3587--3594},
  publisher = {IEEE},
  doi = {10.1109/ITSC58415.2024.10919888},
  abstract = {The development of technologies has prompted a paradigm shift in the automotive industry, with an increasing focus on connected services and autonomous driving capabilities. This transformation allows vehicles to collect and share vast amounts of vehicle-specific and personal data. While these technological advancements offer enhanced user experiences, they also raise privacy concerns. To understand the ecosystem of data collection and sharing in modern vehicles, we adopted the ontology 101 methodology to incorporate information extracted from different sources, including analysis of privacy policies using GPT-4, a small-scale systematic literature review, and an existing ontology, to develop a high-level conceptual graph-based model, aiming to get insights into how modern vehicles handle data exchange among different parties. This serves as a foundational model with the flexibility and scalability to further expand for modelling and analysing data sharing practices across diverse contexts. Two realistic examples were developed to demonstrate the usefulness and effectiveness of discovering insights into privacy regarding vehicle-related data sharing. We also recommend several future research directions, such as exploring advanced ontology languages for reasoning tasks, supporting topological analysis for discovering data privacy risks/concerns, and developing useful tools for comparative analysis, to strengthen the understanding of the vehicle-centric data sharing ecosystem.},
  pdf = {https://www.honda-ri.de/pubs/pdf/5841.pdf}
}

@inproceedings{Shi2022a,
  title = {Uncovering {{Product Vulnerabilities}} with {{Threat Knowledge Graphs}}},
  booktitle = {2022 {{IEEE Secure Development Conference}} ({{SecDev}})},
  author = {Shi, Zhenpeng and Matyunin, Nikolay and Graffi, Kalman and Starobinski, David},
  year = {2022},
  month = {December},
  pages = {84--90},
  doi = {10.1109/SecDev53368.2022.00028},
  abstract = {Threat modeling and security assessment rely on public information on products, vulnerabilities and weaknesses. So far, databases in these categories have rarely been analyzed in combination. Yet, doing so could help predict unreported vulnerabilities and identify common threat patterns. In this paper, we propose a methodology for producing and optimizing a knowledge graph that aggregates knowledge from common threat databases (CPE, CVE, and CWE). We apply the threat knowledge graph to predict associations between threat databases, specifically between products and vulnerabilities. We evaluate the prediction performance based on historical data, using precision, recall, and F1-score metrics. We demonstrate the ability of the threat knowledge graph to uncover many associations that are currently unknown but will be revealed in the future.},
  pdf = {Shi2022 - Uncovering Product Vulnerabilities.pdf}
}

@article{Deng2022,
  title = {Evaluation of {{Cache Attacks}} on {{Arm Processors}} and {{Secure Caches}}},
  author = {Deng, Shuwen and Matyunin, Nikolay and Xiong, Wenjie and Katzenbeisser, Stefan and Szefer, Jakub},
  year = {2022},
  month = {September},
  journal = {IEEE Transactions on Computers},
  volume = {71},
  number = {9},
  pages = {2248--2262},
  issn = {1557-9956},
  doi = {10.1109/TC.2021.3126150},
  abstract = {Timing-based side and covert channels in processor caches continue to be a threat to modern computers. This work shows for the first time, a systematic, large-scale analysis of Arm devices and the detailed results of attacks the processors are vulnerable to. Compared to x86, Arm uses different architectures, microarchitectural implementations, cache replacement policies, etc., which affects how attacks can be launched, and how security testing for the vulnerabilities should be done. To evaluate security, this paper presents security benchmarks specifically developed for testing Arm processors and their caches. The benchmarks are evaluated with sensitivity tests, which examine how sensitive the benchmarks are to having a correct configuration in the testing phase. Further, to evaluate a large number of devices, this work leverages a novel approach of using a cloud-based Arm device testbed for architectural and security research on timing channels and runs the benchmarks on 34 different physical devices. In parallel, there has been much interest in secure caches to defend the various attacks. Consequently, this paper also investigates secure cache architectures using proposed benchmarks. Especially, this paper implements and evaluates secure PL and RF caches, showing the security of PL and RF caches, but also uncovers new weaknesses.},
  pdf = {Deng2021 - Evaluation of Cache Attacks on Arm Processors.pdf}
}

@article{Shi2022,
  title = {Threat {{Modeling Tools}}: {{A Taxonomy}}},
  shorttitle = {Threat {{Modeling Tools}}},
  author = {Shi, Zhenpeng and Graffi, Kalman and Starobinski, David and Matyunin, Nikolay},
  year = {2022},
  month = {July},
  journal = {IEEE Security \& Privacy},
  volume = {20},
  number = {4},
  pages = {29--39},
  issn = {1558-4046},
  doi = {10.1109/MSEC.2021.3125229},
  abstract = {Threat modeling tools allow for identifying weaknesses in a system's design. Yet, understanding conceptual differences between the tools is not trivial. We propose a taxonomy of threat modeling tools, using it to compare several popular ones.},
  pdf = {Shi2021 - Threat Modeling Tools A Taxonomy.pdf}
}

@phdthesis{Matyunin2022,
  title = {Sensor-{{Based Covert Channels}} on {{Mobile Devices}}},
  author = {Matyunin, Nikolay},
  year = {2022},
  month = {July},
  address = {Darmstadt},
  abstract = {Smartphones have become ubiquitous in our daily activities, having billions of active users worldwide. The wide range of functionalities of modern mobile devices is enriched by many embedded sensors. These sensors, accessible by third-party mobile applications, pose novel security and privacy threats to the users of the devices. Numerous research works demonstrate that user keystrokes, location, or even speech can be inferred based on sensor measurements. Furthermore, the sensor itself can be susceptible to external physical interference, which can lead to attacks on systems that rely on sensor data. In this dissertation, we investigate how reaction of sensors in mobile devices to malicious physical interference can be exploited to establish covert communication channels between otherwise isolated devices or processes. We present multiple covert channels that use sensors? reaction to electromagnetic and acoustic interference to transmit sensitive data from nearby devices with no dedicated equipment or hardware modifications. In addition, these covert channels can also transmit information between applications within a mobile device, breaking the logical isolation enforced by the operating system. Furthermore, we discuss how sensor-based covert channels can affect privacy of end users by tracking their activities on two different devices or across two different applications on the same device. Finally, we present a framework that automatically identifies covert channels that are based on physical interference between hardware components of mobile devices. As a result of the experimental evaluation, we can confirm previously known covert channels on smartphones, and discover novel sources of cross-component interference that can be used to establish covert channels. Focusing on mobile platforms in this work, we aim to show that it is of crucial importance to consider physical covert channels when assessing the security of the systems that rely on sensors, and advocate for holistic approaches that can proactively identify and estimate corresponding security and privacy risks.},
  school = {Technical University of Darmstadt, Germany},
  doi = {https://doi.org/10.26083/tuprints-00021975},
  url = {http://tuprints.ulb.tu-darmstadt.de/21975/},
  pdf = {Matyunin - Sensor-based Covert Channels on Mobile Devices.pdf}
}

@inproceedings{Canakci2022,
  title = {{{TargetFuzz}}: {{Using DARTs}} to {{Guide Directed Greybox Fuzzers}}},
  shorttitle = {{{TargetFuzz}}},
  booktitle = {Proceedings of the 2022 {{ACM}} on {{Asia Conference}} on {{Computer}} and {{Communications Security}}},
  author = {Canakci, Sadullah and Matyunin, Nikolay and Graffi, Kalman and Joshi, Ajay and Egele, Manuel},
  year = {2022},
  month = may,
  series = {{{ASIA CCS}} '22},
  pages = {561--573},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3488932.3501276},
  abstract = {Software development is a continuous and incremental process. Developers continuously improve their software in small batches rather than in one large batch. The high frequency of small batches makes it essential to use effective testing methods that detect bugs under limited testing time. To this end, researchers propose directed greybox fuzzing (DGF) which aims to generate test cases towards stressing certain target sites. Different from the coverage-based greybox fuzzing (CGF) which aims to maximize code coverage in the whole program, the goal of DGF is to cover potentially buggy code regions (e.g., a recently modified program region). While prior works improve several aspects of DGF (such as power scheduling, input prioritization, and target selection), little attention has been given to improving the seed selection process. Existing DGF tools use seed corpora mainly tailored for CGF (i.e., a set of seeds that cover different regions of the program). We observe that using CGF-based corpora limits the bug-finding capability of a directed greybox fuzzer. To mitigate this shortcoming, we propose TargetFuzz, a mechanism that provides a DGF tool with a target-oriented seed corpus. We refer to this corpus as DART corpus, which contains only 'close' seeds to the targets. This way, DART corpus guides DGF to the targets, thereby exposing bugs even under limited fuzzing time. Evaluations on 34 real bugs show that AFLGo (a state-of-the-art directed greybox fuzzer), when equipped with DART corpus, finds 10 additional bugs and achieves \$4.03x speedup, on average, in the time-to-exposure compared to a generic CGF-based corpus.},
  isbn = {978-1-4503-9140-5},
  pdf = {Canakci2022 - TargetFuzz.pdf}
}

@inproceedings{Matyunin2019a,
  title = {{{MagneticSpy}}: {{Exploiting Magnetometer}} in {{Mobile Devices}} for {{Website}} and {{Application Fingerprinting}}},
  shorttitle = {{{MagneticSpy}}},
  booktitle = {Proceedings of the 18th {{ACM Workshop}} on {{Privacy}} in the {{Electronic Society}}},
  author = {Matyunin, Nikolay and Wang, Yujue and Arul, Tolga and Kullmann, Kristian and Szefer, Jakub and Katzenbeisser, Stefan},
  year = {2019},
  month = nov,
  series = {{{WPES}}'19},
  pages = {135--149},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3338498.3358650},
  abstract = {Recent studies have shown that aggregate CPU usage and power consumption traces on smartphones can leak information about applications running on the system or websites visited. In response, access to such data has been blocked for mobile applications starting from Android 8. In this work, we explore a new source of side-channel leakage for this class of attacks. Our method is based on the fact that electromagnetic activity caused by mobile processors leads to noticeable disturbances in magnetic sensor measurements on mobile devices, with the amplitude being proportional to the CPU workload. Therefore, recorded sensor data can be analyzed to reveal information about ongoing activities. The attack works on a number of devices: we evaluated 80 models of modern smartphones and tablets and observed the reaction of the magnetometer to the CPU activity on 56 of them. On selected devices we were able to successfully identify which application has been opened (with up to 90\% accuracy) or which web page has been loaded (up to 91\% accuracy). The presented side channel poses a significant risk to end users' privacy, as the sensor data can be recorded from native apps or even from web pages without user permissions. Finally, we discuss possible countermeasures to prevent the presented information leakage.},
  isbn = {978-1-4503-6830-8},
  pdf = {Matyunin2019 - MagneticSpy.pdf}
}

@inproceedings{Matyunin2019,
  title = {Vibrational {{Covert Channels}} Using {{Low-Frequency Acoustic Signals}}},
  booktitle = {Proceedings of the {{ACM Workshop}} on {{Information Hiding}} and {{Multimedia Security}}},
  author = {Matyunin, Nikolay and Wang, Yujue and Katzenbeisser, Stefan},
  year = {2019},
  month = jul,
  series = {{{IH}}\&{{MMSec}}'19},
  pages = {31--36},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3335203.3335712},
  abstract = {In this paper, we examine how acoustic signals in sub-bass and infrasonic range can be used to establish a vibrational covert channel between speaker-equipped computers and mobile devices. We show that typical consumer speakers are capable of producing low-frequency sounds, which are not perceivable by humans. At the same time, we show that producing such sounds by the speaker's woofer inevitably generates slight vibrations of the speaker and the surface where it is located. Being unnoticeable to people, such vibrations can be captured by the accelerometer sensor of a mobile device located on the same surface. Therefore, information can be encoded into low-frequency sounds played by a speaker and received on a mobile device by analyzing the produced vibrations. Note that access to the accelerometer on modern mobile devices does not require any user permissions, making the transmission completely unnoticeable. We evaluate the presented covert channel for different speakers, apply it to several application scenarios, and give an overview of possible countermeasures.},
  isbn = {978-1-4503-6821-6},
  pdf = {Matyunin2019 - Vibrational Covert Channels.pdf}
}

@inproceedings{Matyunin2018a,
  title = {Tracking {{Private Browsing Sessions}} Using {{CPU-based Covert Channels}}},
  booktitle = {Proceedings of the 11th {{ACM Conference}} on {{Security}} \& {{Privacy}} in {{Wireless}} and {{Mobile Networks}}},
  author = {Matyunin, Nikolay and Anagnostopoulos, Nikolaos A. and Boukoros, Spyros and Heinrich, Markus and Schaller, Andr{\'e} and Kolinichenko, Maksim and Katzenbeisser, Stefan},
  year = {2018},
  month = jun,
  series = {{{WiSec}} '18},
  pages = {63--74},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3212480.3212489},
  abstract = {In this paper we examine the use of covert channels based on CPU load in order to achieve persistent user identification through browser sessions. In particular, we demonstrate that an HTML5 video, a GIF image, or CSS animations on a webpage can be used to force the CPU to produce a sequence of distinct load levels, even without JavaScript or any client-side code.These load levels can be then captured either by another browsing session, running on the same or a different browser in parallel to the browsing session we want to identify, or by a malicious app installed on the device. To get a good estimation of the CPU load caused by the target session, the receiver can observe system statistics about CPU activity (app), or constantly measure time it takes to execute a known code segment (app and browser). Furthermore, for mobile devices we propose a sensor-based approach to estimate the CPU load, based on exploiting disturbances of the magnetometer sensor data caused by the high CPU activity.Captured loads can be decoded and translated into an identifying bit string, which is transmitted back to the attacker. Due to the way loads are produced, these methods are applicable even in highly restrictive browsers, such as the Tor Browser, and run unnoticeably to the end user. Therefore, unlike existing ways of web tracking, our methods circumvent most of the existing countermeasures, as they store the identifying information outside the browsing session being targeted.Finally, we also thoroughly evaluate and assess each presented method of generating and receiving the signal, and provide an overview of potential countermeasures.},
  isbn = {978-1-4503-5731-9},
  pdf = {Matyunin2018 - Tracking Private Browsing Sessions.pdf}
}

@inproceedings{Matyunin2018,
  title = {Zero-Permission Acoustic Cross-Device Tracking},
  booktitle = {2018 {{IEEE International Symposium}} on {{Hardware Oriented Security}} and {{Trust}} ({{HOST}})},
  author = {Matyunin, Nikolay and Szefer, Jakub and Katzenbeisser, Stefan},
  year = {2018},
  month = apr,
  pages = {25--32},
  doi = {10.1109/HST.2018.8383887},
  abstract = {Adversaries today can embed tracking identifiers into ultrasonic sound and covertly transmit them between devices without users realizing that this is happening. To prevent such emerging privacy risks, mobile applications now require a request for an explicit user permission, at run-time, to get access to a device's microphone. In this paper, however, we show that current defenses are not enough. We introduce a novel approach to acoustic cross-device tracking, which does not require microphone access, but instead exploits the susceptibility of MEMS gyroscopes to acoustic vibrations at specific (ultrasonic) frequencies. Currently, no permissions are needed to access the gyroscope's data, and the gyroscope can be accessed from apps or even from a web browser. In this manner, gyroscopes in modern smartphones and smartwatches can be used as zero-permission receivers of ultrasonic signals, making cross-device tracking completely unnoticeable to users. We evaluate our approach on several mobile devices using different audio hardware, achieving 10-20bit/s transmission bandwidth at distances from 35cm to 16m in realistic attack scenarios. Finally, we discuss potential countermeasures against the presented attack.},
  pdf = {Matyunin2018 - Zero-permission Acoustic Cross-device Tracking.pdf}
}

@inproceedings{Matyunin2016,
  title = {Covert Channels Using Mobile Device's Magnetic Field Sensors},
  booktitle = {2016 21st {{Asia}} and {{South Pacific Design Automation Conference}} ({{ASP-DAC}})},
  author = {Matyunin, Nikolay and Szefer, Jakub and Biedermann, Sebastian and Katzenbeisser, Stefan},
  year = {2016},
  month = jan,
  pages = {525--532},
  issn = {2153-697X},
  doi = {10.1109/ASPDAC.2016.7428065},
  abstract = {This paper presents a new covert channel using smartphone magnetic sensors. We show that modern smartphones are capable to detect the magnetic field changes induced by different computer components during I/O operations. In particular, we are able to create a covert channel between a laptop and a mobile device without any additional equipment, firmware modifications or privileged access on either of the devices. We present two encoding schemes for the covert channel communication and evaluate their effectiveness.},
  pdf = {Matyunin2016 - Covert channels using mobile device's magnetic field sensors.pdf}
}